As the industry accelerates carbon capture, use, and storage initiatives, modeling innovations for CO2 injection and enhanced oil recovery (EOR) have become critical for optimizing recovery and ensuring secure storage. Recent studies highlight a shift toward data-driven and hybrid approaches that combine computational efficiency with operational practicality.
Paper SPE 227168 introduces a deep-learning framework using the temporal fusion transformer (TFT) to optimize sequestration efficiency and oil recovery. Unlike conventional machine learning, TFT captures temporal dependencies and uncertainty, enabling dynamic optimization under varying reservoir conditions. This approach supports real-time decision-making while reducing computational demands compared with full-physics simulations.
Meanwhile, Paper SPE 221978 focuses on fast predictive models for mature oil fields, using surrogate techniques such as artificial neural networks and autoregressive models. These proxies significantly accelerate scenario evaluation, making them ideal for screening and feasibility studies where time and resources are limited. The inclusion of uncertainty quantification enhances reliability for planning under variable conditions.
A third reviewed paper, IPTC 25000, addresses well-network design for CO2 EOR and storage through an integrated modeling approach that couples reservoir simulation with network-optimization algorithms. Applied at field scale, this method improved sweep efficiency and reduced CO2-breakthrough risk, demonstrating scalability for heterogeneous reservoirs.
These advances mark a paradigm shift toward intelligent, integrated modeling frameworks in EOR. Future efforts should aim to unify deep learning, surrogate modeling, and network optimization into holistic platforms, enabling end-to-end optimization across reservoir and infrastructure scales. Such innovations will be pivotal in achieving dual objectives: maximizing oil recovery and ensuring secure CO2 storage in the energy-transition era.
Summarized Papers in This January 2026 Issue
SPE 227168 Deep-Learning Technique Optimizes Sequestration, Oil Production in CCUS Projects by Ahmed Wagia-Alla, SPE, Mohamed Alghazal, and Turki Alzahrani, SPE, Saudi Aramco.
SPE 221978 Fast Predictive Models Developed for CO2 EOR and Storage in Mature Oil Fields by Yessica Peralta, Ajay Ganesh, and Gonzalo Zambrano, SPE, University of Alberta, et al.
IPTC 25000 Integrated Well-Network-Design Mode Developed for CO2 EOR and Storage by Zangyuan Wu, PetroChina, CNPC, and China University of Petroleum; Yongliang Tang, PetroChina and CNPC; and Liming Lian, CNPC, et al.
Recommended Additional Reading
SPE 221850 The First Application of Quantum Computing Algorithm in Streamline-Based Simulation of Waterflooding Reservoirs by Xiang Rao, Yangtze University
SPE 227695 CO2 Flooding Optimization Using Artificial Neural Networks: Enhancing Oil Recovery and Carbon Sequestration by N.A. Almakki, University of Khartoum, et al.
SPE 224577 Leveraging Machine Learning To Model Hydrocarbon/CO2 Solubility Behavior by Seyed Mehdi Alizadeh, Australian University, et al.
Luky Hendraningrat, SPE, is a senior scientist in reservoir technology at Petronas. He holds a doctoral degree in enhanced oil recovery (nanoparticles) from the Norwegian University of Science and Technology. Hendraningrat has more than 21 years of oil and gas experience. His research interests are improved/enhanced oil recovery, pressure/volume/temperature analysis, and reservoir modeling. Hendraningrat has published more than 75 technical papers. He has volunteered on technical program committees for multiple SPE events and is a recipient of the 2024 SPE Regional Service Award.
